Biology subjects

Feng, T.

Publications and source records attributed to Feng, T..

At least 19 recordsLinked to original sources

bioRxiv2026

Aberrant microglial responses shape hypothalamic circuits in anorexia nervosa

Multimodal data indicates that microglia contribute to the pathophysiology of anorexia nervosa (AN). Here, we investigated microglial modulation of hypothalamic circuits, key regulators of energy balance, hypothesising an implication in the prolonged starvation and low body weight of AN. First, we generated microglia as well as hypothalamic and cortical neurons from patient-derived induced pluripotent stem cells (iPSCs) to discover upregulation of synapse-related genes in hypothalamic neurons, reduced microglial uptake of hypothalamic synaptic structures, and a microglial unresponsivness to the satiety hormone glucagon-like peptide-1. In an AN animal model (anx/anx mouse), spatial transcriptomics indicated hypothalamic microglial activation and disrupted microglia-synapse signaling. Despite an increased microglia density in both the arcuate nucleus (ARC) and the dorsomedial hypothalamus region (DMH), microglia displayed a decreased per cell uptake of synaptic material in Arc. Together, these data suggest that microglial responses shape hypothalamic circuits with possible implications for the maintained negative energy balance of AN.

neuroscience↗

bioRxiv2026

GO, NO-GO: A SENSITIVE PERIOD FOR DEVELOPING SEX-SPECIFIC VISUALLY-GUIDED PREY PURSUIT-PREDATOR AVOIDANCE TRADE-OFF STRATEGIES

In natural environments, animals face a trade-off between foraging and staying vigilant against predators, a conflict known to shape sensory processing and behavior through natural selection. Even within a species, this trade-off may be resolved differently depending on age, sex, or life history. To identify the neural mechanisms underlying visually-guided pursuit-avoidance trade-offs, we studied how mice responded to a sudden overhead threat while pursuing a moving, prey-like target near the ground. This paradigm let us quantify orienting decisions in adolescent and adult mice of both sexes. We found that adolescence is a key period when male and female mice begin to diverge in how they resolve this conflict. Manipulating the value of the pursued target during adolescence versus adulthood revealed that adolescence is also a sensitive period for shaping adult escape-to-shelter versus continue to approach target behavior. Notably, hunting experience gained specifically during adolescence caused males, but not females, to shift strategy: experienced adolescent males tended to uniquely shift towards a "no-go" (stay-near-Prey) strategy. All the other groups instead favored an active "go" strategy, repeatedly shifting between approaching target and running to shelter. All mice showed this oscillation between approaching target and escaping to a shelter to some degree, but hunting experience during adolescence most significantly shifted this balance between approach and escape for males.

neuroscience↗

bioRxiv2026

Systematic benchmarking of zero-shot utility and robustness in single-cell transcriptomic foundation models

Single-cell foundation models (scFMs) have emerged as powerful representation learning approaches for single-cell transcriptomics. However, the utility and robustness of their pretrained representations across diverse analytical tasks and data conditions remain insufficiently characterized, particularly in zero-shot settings without task-specific fine-tuning. Here, we systematically analyze zero-shot performance of single-cell transcriptomic representations across 20 methods, 6 downstream tasks and 1,607 datasets comprising nearly 21.8 million cells. We evaluate model behavior along three complementary dimensions: utility on original datasets, robustness to controlled changes in dataset structure, and exploratory associations between dataset characteristics and performance variation. Our results show that scFM performance is strongly task dependent, with no single method consistently outperforming others across cell- and gene-level analyses. Notably, high utility on original datasets did not necessarily translate into robustness under structural perturbations, and several top-ranking methods were sensitive to changes in cell number, gene number, class composition, class imbalance, and batch complexity. Conventional statistical and task-specific methods remained competitive in several settings, while greater computational cost did not consistently correspond to better performance. Driver analyses further identified task-specific associations between performance and dataset characteristics, including cell-type complexity, train-test class overlap, batch number, and regulatory target-set size. Together, these findings show that the zero-shot utility and robustness of pretrained scFM representations depend jointly on analytical task and dataset structure. Our study provides a practical basis for context-aware representation selection and underscores the importance of evaluating structural robustness alongside utility when developing and applying scFMs.

bioinformatics↗

bioRxiv2026

A distributed neural architecture of sustained affect across external and internal experience

AO_SCPLOWBSTRACTC_SCPLOWSustained affect shapes well-being, yet its neural architecture across externally elicited and internally generated experience remains unclear. Using whole-brain functional connectivity during minutes-long naturalistic movie viewing, we derived positive and negative affective experience signatures and their underlying neural architecture. These signatures predicted valence-specific affective intensity and generalized to independent movie-viewing data and internally generated affect, discriminating sad memory and rumination from neutral distraction while tracking subjective experience. Importantly, their expression showed little relation to vigilance or cognitive demand. Characterization of these signatures revealed coherent community structure and a shared distributed backbone, alongside valence-preferential components, consistent with a partially separable architecture. Extending beyond experimentally evoked states, in four resting-state depression cohorts, these signatures distinguished patients from controls with reduced positive and elevated negative signature expression, and predicted symptom burden and anhedonia. These findings identify a generalizable distributed architecture bridging external and internal affective experience and extending to clinically relevant affective dysregulation.

neuroscience↗

bioRxiv2026

A human-in-the-loop explanation framework for morphologically transparent AI predictions from whole-slide images

Deep learning models enable the prediction of clinical endpoints from whole-slide images (WSIs), but many such models function as "black boxes", lacking transparency about whether and which histomorphological patterns drive their predictions, hindering interpretability and clinical adoption. Here we propose a human-in-the-loop explanation framework, MorphoXAI, which provides both local and global interpretability for deep learning models by incorporating human-expert interpretations. At the global level, it reveals the histomorphological patterns on which the model consistently relies to distinguish between classes of WSIs, as well as the patterns associated with confusion between classes. At the local level, it indicates which of these patterns are used in the prediction of an individual WSI and which regions within the slide correspond to such patterns. We validated our method on a deep learning model trained for ovarian tumor histologic subtype prediction. The results show that our framework generates explanations that accurately reflect the histomorphology underlying the models predictions at both global and local levels. For interpretability and clinical utility in diagnostic contexts, human evaluation results showed that our explanations were easy to interpret, rich in diagnostic features, and directly helpful for diagnostic decision-making, thereby enhancing pathologist-AI collaboration. Our work highlights that unifying global and local explanations and grounding them in expert-interpreted morphology enhances the interpretability and verifiability of deep learning models, thereby facilitating the transparent deployment of such models in clinical practice.

pathology↗

bioRxiv2026

Enhancing Lipid Detection and Spatial Accuracy in Carotid Plaques Using Mass Spectrometry Imaging Techniques

Matrix-assisted laser desorption/ionisation mass spectrometry imaging (MALDI-MSI) is a powerful technique for studying lipid distribution in carotid plaques, key to understanding atherosclerosis. This study aimed to improve sample preparation for MALDI-MSI-based spatial lipidomics of carotid plaques by improving both matrix application and tissue handling. Human carotid plaques were collected from endarterectomy patients with ethical approval and sectioned at 10 {micro}m thickness for MALDI-MSI. We compared eight sample preparation methods, including hydroxypropyl methylcellulose-polyvinylpyrrolidone (HPMC-PVP) embedding media and Cryofilm-type IMS(R) to provide support and maintain tissue structural integrity during sectioning. Methods were assessed for signal intensity, lipid diffusion, lipid coverage, tissue morphology, and image co-registration which each criterion scored from 1-3. Cryofilm-based methods scored highest for preserving tissue morphology and minimising folding artifacts (2.9-3.0) but were limited in co-registration (2.0) due to reliance on adjacent sections. Sublimation methods generally produced greater lipid coverage with reduced lateral diffusion, while automated sprayer methods scored higher in signal intensity/sensitivity (3.0) but had increased lipid delocalisation, particularly for highly hydrophobic species such as triacylglycerols and sterols. The results highlight clear trade-offs between tissue structural preservation, lipid detection sensitivity, and spatial integrity in MALDI-MSI. Because spatial integrity cannot be compromised for imaging lipids in carotid atherosclerotic plaques, Cryofilm combined with sublimation offers a clear advantage. This work strengthens MALDI-MSI workflows enabling more precise spatial mapping and deeper biological interpretation of atherosclerotic lipid distributions.

molecular biology↗

bioRxiv2026

MOV10 inhibits SADS-CoV replication by enhancing TRIM24-mediated K63-linked TRAF3 ubiquitination and this inhibition is antagonized by viral N protein

The global outbreak of SARS-CoV-2 has resulted in a renewed focus on coronaviruses with the potential for cross-species transmission and mutation risks. In particular, swine acute diarrhea syndrome coronavirus (SADS-CoV), potentially originating from intermediate horseshoe bat (Rhinolophus affinis), is a novel coronavirus that causes acute diarrhea, vomiting, and high mortality in suckling piglets. Innate immunity plays a crucial role in defending the host against invading pathogens. Previous studies have identified several host factors that inhibit SADS-CoV replication through innate immunity; however, research on SADS-CoV remains insufficient compared to that on other coronaviruses. In this study, the host factor Moloney leukemia virus 10 protein (MOV10) inhibited SADS-CoV replication by promoting interferon (IFN) production. Mechanistically, MOV10 exerted its antiviral effect primarily through its N-terminal domain, which regulated the TNF receptor-associated factor 3 (TRAF3)-mediated innate immune pathway. Specifically, we reveal that MOV10 enhanced tripartite motif-containing 24 (TRIM24)-mediated K63-linked ubiquitination of TRAF3, thereby promoting IFN production and inhibiting SADS-CoV replication. Furthermore, we identified the strategies by which SADS-CoV evaded innate immunity. The viral N protein inhibited the antiviral effect of MOV10 by disrupting interaction between MOV10 and TRAF3, and by promoting K48-linked polyubiquitination of TRAF3, leading to TRAF3 degradation via the ubiquitin-proteasome pathway. Collectively, our findings reveal the role of MOV10 in antiviral immunity and how SADS-CoV evades host immune defense.

molecular biology↗

bioRxiv2025

Effects of Combined Viral and Bacterial Vaccine on Production Performance in Broiler Chickens

This study formulated a novel inactivated vaccine containing Newcastle disease virus (Lasota strain), avian influenza virus (BZ strains), APG, adenovirus and ORT. The vaccine was administered to 1-day-old and 7-day-old fast-growing broiler chickens. Experimental results showed that the vaccine reduced the feed conversion ratio and induced high levels of antibodies against Newcastle disease (Lasota strain), avian influenza (BZ strains), and avian rhinotracheitis. Immunization at 1 day of age had a significant impact on the production of antibodies against adenovirus. Challenge tests indicated that the vaccine partially protected the broiler chickens against Avibacterium paragallinarum types A, B, and C in broilers. Post-mortem examination of the injection sites showed good absorption with no significant unabsorbed oil droplets or hyperplasia in both 1-day-old and 7-day-old immunized chickens. This study demonstrated that combined viral and bacterial vaccine enhanced the feed conversion ratio in broiler chickens.

microbiology↗

bioRxiv2025

Immunogenicity of different types of adjuvants in bacterial vaccine

This study aims to investigate four common trivalent avian infectious coryza (AIC) vaccines available on the market. It detects the antibody levels of three serotypes at 7 days and 14 days after the first immunization, and at 30 days, 90 days, and 180 days after the second immunization using HI (Hemagglutination Inhibition Test) and ELISA (Enzyme-Linked Immunosorbent Assay). Additionally, the levels of cytokines including IFN-{gamma} (Interferon-Gamma), IL-2 (Interleukin-2), and IL-3 (Interleukin-3) are measured. Furthermore, the protective rates of four of these vaccines are verified 180 days later. It demonstrated that the immunogenicity of mineral oil adjuvanted vaccine is higher than that of Aluminum adjuvanted. The animals should be boosted with another booster immunization six month post the first immunization to protect them from infectious coryza.

microbiology↗

bioRxiv2025

Characteristics of reproductive tract microbiota in health and disease

The symbiotic relationship between the microbiota and the host is crucial to host health. The reproductive tract microbiota in health and disease was examined by next-generation sequencing for the first time in this study. A total of 10 healthy and 30 diseased samples were collected for microbiota analysis. Species distributions, alpha diversity, beta diversity, and functional gene prediction were analyzed. Riemerella anatipestifer, Escherichia coli, Avibacterium paragallinarum, Ornithobacterium rhinotracheale, Enterococcus faecalis, and Acinetobacter baumannii were first identified as pathogenic bacteria at higher abundance in the reproductive tract of the laying hens. The diseased group exhibited dysbacteriosis and a different species-abundance cluster compared with the healthy control. Alpha-diversity analysis showed that the ACE index in the control group was significantly higher than in the diseased group. Beta diversity analysis indicates that samples from the healthy group had a more similar composition than those from the diseased group. Moreover, functional gene prediction analysis indicated that diseased groups showed a high level of gram-negative, facultative anaerobic, and potentially pathogenic bacteria. The pathogenic bacteria identified in this study are helpful for vaccine development and disease treatment. In conclusion, next-generation sequencing is an effective method for analyzing bacterial communities to support health assessment and disease diagnosis.

microbiology↗

bioRxiv2025

CARP Protects Against Doxorubicin-induced Cardiotoxicity through NRF1-driven Mitochondrial Homeostasis

AimsDoxorubicin (DOX), a highly effective anthracycline chemotherapeutic agent, is limited by its dose-dependent cardiotoxicity. This study investigates the cardioprotective mechanisms of cardiac adriamycin responsive protein (CARP) and its underlying mechanisms in DOX-induced cardiotoxicity (DIC). Methods and resultsCardiac-specific CARP transgenic and wild-type mice were subjected to a DIC model. Cardiac function was assessed via echocardiography, histopathology, and transmission electron microscopy (TEM). In vitro, DOX-treated cardiomyocytes overexpressing CARP were analyzed for oxidative stress (ROS levels), mitochondrial function (mtDNA copy number etc.), and mitochondrial-related proteins (Western blot). CARP-NRF1 interaction was validated by co-immunoprecipitation (Co-IP), and NRF1 siRNA knockdown was performed to assess CARPs role in mitochondrial homeostasis. CARP overexpression markedly alleviated DOX-induced cardiac dysfunction and mitochondrial damage, restoring mitochondrial dynamics and mitophagy. Mechanistically, CARP directly interacted with NRF1, and NRF1 knockdown ameliorated CARP-mediated cardioprotection in DIC. ConclusionCARP safeguards against DIC by maintaining mitochondrial homeostasis via NRF1 signaling, positioning it as a promising therapeutic target for DOX-induced cardiomyopathy.

molecular biology↗

bioRxiv2025

Transforming a Fragile Protein Helix into an Ultrastable Scaffold via a Hierarchical AI and Chemistry Framework

The rational design of proteins that maintain structural integrity under concurrent thermal, mechanical, and chemical stress remains a challenge in molecular engineering. We present a hierarchical framework that transforms an -helical domain into an ultrastable scaffold by integrating AI-guided design with foundational chemical principles. This approach progresses from global architectural reinforcement, using multiple AI tools to create a stabilized four-helix bundle, to local chemical tuning, where AlphaFold3 guides the installation of salt bridges and metal-coordination motifs. A computational pipeline using physics-based screening such as molecular dynamics (MD) simulations efficiently distilled millions of designs into a minimal candidate set. The resulting -helical proteins exhibit unprecedented multi-axis stability, with mechanical unfolding forces exceeding 200 pN, thermal resilience >100 {degrees}C, and high resistance to chemical denaturants. By systematically dissecting the contributions of hydrophobic packing, electrostatics, and metal coordination, we establish a general blueprint for imparting extreme robustness. This work bridges AI-driven structural generation with chemical precision, advancing the creation of durable proteins for mechanistic studies and synthetic biology.

biophysics↗

bioRxiv2025

The Lettuce Expression Browser: from lab to LEB

Lettuce (Lactuca sativa L.) is an economically important leafy vegetable within the Asteraceae family, cultivated worldwide across diverse agricultural systems. Recent advances in genomic and transcriptomic resources have positioned lettuce as a promising model system for functional genomics in the Asteraceae. However, currently available gene expression datasets lack comprehensive tissue-specific resolution, primarily focus on a single cultivar and are not visualised in an interpretable manner, limiting their utility for broader genetic and physiological studies. To bridge this gap, we developed the Lettuce Expression Browser (LEB), a publicly available platform providing high-resolution gene expression maps across various organs, tissues and developmental stages in both cultivated and wild lettuce species. The LEB integrates transcriptomic data from finely dissected seedlings, shoot tissues at various developmental stages and seedlings subjected to abiotic stresses (salt and far-red), visualised using the ggPlantmap R package. This platform offers an intuitive interface for exploring gene expression patterns and serves as a valuable resource for those studying lettuce development, stress responses, and evolutionary genomics. The LEB is hosted on the LettuceKnow Web Portal (https://lettuce.bioinformatics.nl) and can be expanded to include additional datasets, enhancing its role as a key tool for lettuce research and crop improvement. Significance statementThe Lettuce Expression Browser (LEB) provides the first high-resolution gene expression atlas for both cultivated and wild lettuce species. This open-access resource enables detailed exploration of gene activity across development stages and stress conditions, advancing functional genomics in the Asteraceae family.

plant biology↗

bioRxiv2025

Procrastination partly reflects an evolutionary byproduct of non-planning impulsivity

Procrastination has immediately visible repercussions on health and survival resilience, yet shows stably heritable and remains increasingly pervasive across human societies. Despite a paradox, this behavior is theoretically explained to represent a byproduct of evolutionary advantages underlying impulsivity, yet not deciphered well by scientific evidences. After adjusting psychometric endogeneity, we demonstrate the unique predictive roles of non-planning impulsivity (NPI) during late adolescence and early adulthood uniquely predicts procrastination in later adulthood in a twin cohort (N = 154). This association was further replicated in two independent cohorts (N = 327, N = 1,543). Using AE models, in conjunction of single-paper meta-analytic synthesis (N = 3,656 twin pairs), we observed significant shared genetic contributions underlying this NPI-procrastination association (rg = 0.51, 95% CI: 0.18 - 0.84). Beyond to the phenotypic heritability, employing a Genome-Wide-Association Study (GWAS), six NPI-procrastination overlapping SNPs are identified, functionally accounting for neural dysregulation. Thus, leveraging neurodevelopmental normative modelling (N = 37,407), online meta-analytic estimations (k = 198, loci = 5,855) and seed-based d mapping estimates (N = 893), cortical deviations in the left dorsolateral prefrontal cortex (DLPFC) - the brain region showing highest probabilistic overlap mapping NPI to procrastination, partly explains their shared genetic variants, but are substantially independent in genetic contribution. Mendelian Randomization analysis finally indicates causal roles of NPI and procrastination both, to DLPFC deviations. Our findings empirically clarified this theory that procrastination partly derives from NPI as an evolutionary byproduct indeed, but is still unique in neurogenetic entities. Brief summaryProcrastination is a puzzling human behavior that compromises survival-relevant outcomes yet remains both widespread and heritable. Although theorized as a byproduct of impulsivitys evolutionary advantages, empirical support for this account has been limited. Here, we provide converging evidence across psychometric, genetic, and neuroimaging modalities to show that non-planning impulsivity during late adolescence and early adulthood uniquely predicts later procrastination, and that the two traits share significant genetic overlap. We further identify specific genetic variants and morphological deviations in the DLPFC that link, but also partially dissociate, their biological pathways. These findings clarify the evolutionary and neurogenetic architecture of procrastination and underscore its partial derivation from impulsivity alongside distinct developmental origins.

neuroscience↗

bioRxiv2025

A study of different pre-treatment methods for cigarettes and their aroma differences

The study innovatively utilized a combination of smoking machine and Cambridge filter to effectively extract the aroma compounds generated during smoking of cigarettes, and employed three different pretreatment methods to obtain optimal cigarette pretreatment procedures. In addition, the aroma components of different cigarettes were identified by different instruments such as gas chromatography-mass spectrometry. Combined with the calculation of Odor Activity Values (OAV), PCA analysis and sensory evaluation were performed on the five types of cigarettes to reveal differences and similarities in their aroma profiles.

biochemistry↗

bioRxiv2025

Sequencing-based Spatial Transcriptomics with scRNA-seq Sensitivity

The advent of spatial transcriptomics has dramatically expanded our ability to study the vast network of cell-cell interactions at the molecular level in tissue. Among current methods, sequencing-based approaches have great potential in discovering because of its unbiased capture. In the last couple of years, the spatial resolution for the capture addresses has been significantly improved from 100 m to <1 m, well below the size of a mammalian cell. However, the capture efficiency has always been a pain point, [~]one order of magnitude lower than that of single cell RNA sequencing (scRNA-seq). The low capture efficiency limits the depth and breadth of its applications in the study of complex biological systems and diseases. Here, we introduce Salus Spatial transcriptomic system (Salus-STS), which provides [~]1 m capture resolution and a capture efficiency [~]1 order of magnitude higher than other current methods. Analysis with sub-cellular resolution becomes practical for sequencing-based spatial transcriptomics.

molecular biology↗

bioRxiv2025

Phylogenomic synteny analysis tracks conserved ancient polyploid-derived triplicated genomic blocks across Asteraceae genomes

The Asteraceae (Compositae) is the largest flowering plant family, ubiquitous in most terrestrial communities, and morphologically hyper-diverse. An ancient whole genome triplication (paleo-hexaploidization) occurred at approximately the same time as the evolutionary innovation and adaptive radiation of the family during the middle Eocene. Despite its importance, the genomic contents arising from this triplication have yet to be tracked in context of the Asteraceae genome evolution. We applied a synteny oriented phylogenomic analysis of 21 Asterales genomes and to study the paleo-hexaploidization and its consequences to gene, trait, and genome evolution. We identified 15 ancestral linkage groups (ALGs) that date back to the common diploid ancestor of all Asteraceae. Each of these groups was triplicated, resulting in 45 genomic blocks (3x15), which serve as the foundation for cross-family analyses. We demonstrate the complex evolutionary dynamics of the 45 genomic blocks across the Asteraceae phylogeny. We found that modern genomes are genetic mosaics of three progenitor genomes by extensive genomic exchange, chromosomal shuffling and gene fractionation. 157 genes retained three paleo-hexaploid derived syntenic paralogs across most Asteraceae species. Transcription factors (TFs) and auxin-related genes are significantly overrepresented in the conserved triplets, and expression of the paleo-hexaploidy paralogs is spatiotemporally differentiated. These genes are involved in the development of floral capitulum, a remarkable morphological innovation of the family. The discovery of conserved triplicated genes can direct further study to understand the evolutionary innovation, and the synteny-phylogenomic framework and ALGs provide a comparative framework to characterize newly sequenced Asteraceae genomes.

genomics↗

bioRxiv2024

Dissociable neurofunctional and molecular characterizations of reward and punishment sensitivity

While the hyper-and hypo-reward or punishment sensitivities (RS, PS) have received considerable attention as prominent transdiagnostic features of psychopathology, the lack of an overarching neurobiological characterization currently limits their early identifications and neuromodulations. Here we combined microarray data from the Allen Human Brain Atlas with a multimodal fMRI approach to uncover the neurobiological signatures of RS and PS in a discovery-replication design (N=655 participants). Both RS and PS were mapped separately in the brain, with the intrinsic functional connectome in the fronto-striatal network encoding reward responsiveness, while the fronto-insular system was particularly engaged in punishment sensitivity. This dissociable functional connectome patterns related to RS and PS were also specific in differentiating decisions driven by social or monetary reward and punishment motivations. Further imaging transcriptomic analyses revealed that functional connectome variations for RS and PS were associated with topography of specific gene sets enriched in ontological pathways, including synaptic transmission, dopaminergic metabolism, immune response and stress adaptation. On the neurotransmitter level, the serotonin neuromodulator was identified as a pivotal hub regulating the intrinsic functional connectome patterns of RS and PS, with this process critically dependent on its interactions with dopaminergic, opioid and GABAergic systems. Overall, these findings indicate dissociable neural connectome mapping of RS and PS and highlight their linkage with transcriptomic profiles, which may offer valuable insights into the treatment evaluation for symptomatology relevant to reward/punishment processing deficits.

neuroscience↗